Home/ Questions/Q 4337232
In Process

The Archive Base Latest Questions

Editorial Team
  • 0
Asked: 2026-05-21T10:54:57+00:00

I’m currently writing my own parser for a fictional Assembly language. The instructions are

  • 0

I’m currently writing my own parser for a fictional Assembly language. The instructions are very similar to any normal assembly instruction:

[INSTRUCTION] [OP]*

where op can be 0-3 operands. I want to be able to use an expression that matches this. This is being written in C++ with boost::regex. I myself am a regexp noobie, trying to understand the boost documentation of what each symbol does.

Now, I already have an expression that can match 0-3 operands like so:

Sample Instructions:
    MOVI 8 10
    ADDI 8 8 10
    NOP
    BNEZI -1

Expression: ^([a-z]+)( ([-,0-9]+))*

However, I can’t create a suitable expression that handles the same instructions when comma-delimited:

Sample Instructions:
    MOVI 8, 10
    ADDI 8, 8, 10

This is really tripping me up. I tried rewriting my expression like so:

^([a-z]+)( ([-,0-9]+))*(, ([-,0-9]+))*

This looks to be extremely green, poor regexp. It also isn’t working correctly. I was thinking of using a recursive expression, but I looked at the documentation and I might as well scribble “overkill” on my forehead.

I realize I could just format the line to take out all the commas, but I would rather like to be able to write and understand a regexp expression first, then do it the easy way. Any help would be appreciated.

Share

Leave an answer

You must login to add an answer.

Need An Account,

1 Answer

  1. Editorial Team

    A string like:

    ADDI 8, 8, 10

    can be matched by a regex like this:

    [a-zA-Z]+[ \t]+-?[0-9]+([ \t]*,[ \t]*-?[0-9]+)*

    A (short) explanation:

    [a-zA-Z]+   # match an instruction
[ \t]+      # match one or more spaces or tabs
-?[0-9]+    # match an integer value with an option minus sign in front of it
(           # open group 1
  [ \t]*    #   match zero or more spaces or tabs
  ,         #   match a comma
  [ \t]*    #   match zero or more spaces or tabs
  -?[0-9]+  #   match an integer value with an option minus sign in front of it
)*          # close group 1, and repeat it zero or more times

    Having said all that, I must agree with dmckee’s comment: a proper parser is the way to go, even if this is just a fictional language you’re parsing.

    • 0